An approach to model coal combustion process to predict and minimize unburned carbon in bottom ash of a large-capacity pulverized coal-fired boiler used in thermal power plant is proposed. The unburned carbon characteristic is investigated by parametric field experiments. The effects of excess air, coal properties, boiler load, air distribution scheme, and nozzle tilt are studied. An artificial neural network (ANN) is used to model the unburned carbon in bottom ash. A genetic algorithm (GA) is employed to perform a search to determine the optimum level process parameters in ANN model which decreases the unburned carbon in bottom ash.